Nutrition Dropoff Protocol

SHARE Wave 11 in the Czech Republic

1. Background and Aims

Diet is a key, modifiable determinant of healthy ageing, affecting cardiometabolic risk, frailty, sarcopenia, cognitive decline, and disability trajectories. However, robust diet-health analyses in ageing cohorts require dietary exposure measures that are (i) sufficiently detailed for nutrient- and food-group analyses, (ii) comparable across countries and survey waves, and (iii) feasible in older populations with heterogeneous digital skills and health constraints.

SHARE is uniquely positioned to address diet-related research questions in later life because it combines repeated, harmonized measures of health, function, socioeconomic status, and social networks across many European countries. Yet, diet has historically been under-measured in SHARE relative to its importance for life-course epidemiology and to other surveys (ELSA, NICOLA, HRS). Implementing a Czech nutrition dropoff as a country-specific sub-project linked to SHARE Wave 11 creates high scientific value — it enables integration of dietary exposures with rich SHARE phenotyping and supports cross-country harmonisation by aligning with instruments already used in other ageing studies.

The primary research aims of the nutrition dropoff are:

  1. To quantify key dietary exposures in adults aged 50+ in the Czech SHARE sample, including foods, nutrients, and dietary patterns, and to link these exposures with SHARE’s longitudinal health, functional, and socioeconomic outcomes from a life-course perspective.
  2. To enhance cross-study and cross-country comparability of dietary data within ageing research by aligning nutritional outputs with established European population studies and harmonized food and nutrient constructs.
  3. To improve the validity and interpretability of diet-aging associations by addressing known sources of measurement error in dietary assessment.

2. Objectives and content of the nutrition dropoff

Including the collection of newly developed survey measures of nutrition in SHARE data provides important scientific value as it opens up multiple research possibilities:

  1. Diet-ageing trajectories: modeling associations between dietary exposures and longitudinal changes in functional status, frailty, cognition, morbidity, and health service use (leveraging repeated SHARE measures).
  2. Dietary patterns and inequality: analyzing dietary patterns/scores (e.g., PCA-derived patterns, cluster solutions, Mediterranean-style or other diet indices) in relation to socioeconomic position, social networks, and regional variation within an ageing population.
  3. Methodological research: quantifying and correcting measurement error (under-reporting, selective non-response, mode effects) and evaluating feasibility of digital dietary assessment in older adults, including mixed-mode solutions.

The nutrition dropoff is based on existing, validated, and internationally used nutrition instruments:

  1. Food Frequency Questionnaire (EPIC-Norfolk FFQ; EPIC-based) - captures habitual intake through a structured list with standard frequency categories and portion conventions, suitable for long-term exposure ranking and trend comparisons.
  2. 24-hour dietary recall (Intake24) - captures detailed day-level intake (one weekday and weekend day) with automated nutrient analysis, enabling meal-pattern outputs and more granular food group exposure estimates; validation evidence supports its use in large studies as a scalable alternative to interviewer-based 24HDR.

The nutrition dropoff is designed as a pilot study to be used in other SHARE countries. Both FFQ and 24HDR tools are open-source, free questionnaires with free software for nutrition coding.

3. Nutrition Dropoff Design

The nutrition dropoff in the Czech Republic is designed as a country-specific sub-project of Wave 11 of SHARE. SHARE is a multidisciplinary and cross-national panel database of micro data on health, socio-economic status and social and family networks. For details see www.share-eric.eu.

Responsible parties

SHARE-ERIC
Chausseestrasse 111, 10115 Berlin
(hereafter SHARE-ERIC)

SHARE-CZ Country Team
Radim Bohacek (Country Team Leader of SHARE-CZ)
Tereza Vagnerova, PhD (Nutrition Expert at SHARE-CZ and the Department of Geriatrics and Internal Medicine, First Faculty of Medicine, Charles University)
(hereafter SHARE-CZ).

Sample size and target population

The sample size for the nutrition dropoff will include all refresher and panel respondents in the Czech Republic who have completed the main interview in wave 11. The participation in the nutrition dropoff is voluntary. Should a respondent refuse to participate in the nutrition dropoff, no questionnaires will be provided.

Based on ELSA and NICOLA nutrition dropoffs, the expected consent rate is 80% of the target population. The expected number of respondents who will complete the nutrition dropoff is 3,300.

Time Schedule

The nutrition dropoff will be tested in the field rehearsal for wave 11. The nutrition dropoff main data collection will be collected as a dropoff questionnaire in wave 11.

Questionnaires

The nutrition dropoff consists of two nutrition questionnaires which will be adapted to Czech localization by SHARE-CZ. Both FFQ and 24HDR tools are open-source, validated nutrition questionnaires used in scientific projects, developed at the UK by Newcastle University and maintained at the MRC Epidemiology Unit at the University of Cambridge and Monash University (Australia).

Localization into Czech environment

The SHARE-CZ is responsible for producing a harmonized versions of FFQ and 24HDR questionnaires in Czech. Localization of questionnaires, tools, food database, and materials will be provided by SHARE-CZ with collaboration of the MRC Unit at Cambridge University. The MRC Unit provides support for installation of localization toolkit for questionnaire coding. All costs related to localization and translation will be covered by SHARE-CZ budget.

4. The Nutrition Dropoff Questionnaire Content

The dropoff questionnaire consists of two parts, the Food Frequency Questionnaire (FFQ) and the 24 hour dietary recall questionnaires (24HDR). For the questionnaires see Appendix.

1. Food Frequency Questionnaire (EPIC-Norfolk FFQ)

The EPIC-Norfolk Food Frequency Questionnaire (FFQ) consists of an 8-page food list containing commonly and less commonly consumed foods, categorized into 10 groups. In the first main part, the questionnaire contains a list of commonly and less commonly consumed foods. For each item on the list, participants are asked to indicate their usual frequency of consumption choosing from nine frequency categories. The categories range from “never or less than once/month” to “6+ times per day”. The servings are specified in terms of units or common portions (e.g. one apple, one slice of bread) or household measures (e.g. glass, cup, spoon). An average portion size was assigned to each questionnaire item (portion size is the same for all respondents).

Part 2 asks a set of additional questions including those on the type and brand of breakfast cereal, the type and quantity of milk consumed, the type of fat used in cooking and the amount of visible fat on meat consumed. The questions relating to fat are linked to relevant items in part one and help to estimate the total fat and fatty acid consumption.

2. 24h Recall Questionnaire (24HDR)

The 24HDR recall consists of one for a weekday and one for a weekend day. For the 24-hour dietary assessment will be collected using the dietary recall memory cards from which the nutrition information will be entered into Intake24, an open-source dietary self-completion web-based assessment tool developed by Newcastle University and maintained by the MRC Epidemiology Unit at the University of Cambridge. Two identical memory cards are filled separately by the respondent. Each questionnaire is divided into several day periods for which the respondent writes down the food and drinks she or he consumed. The information from the memory card is filled by the respondent to the web-based Intake24 tool, or sent in the paper form to the survey agency where it is filled on behalf of the respondent.

5. Fieldwork Procedures

Sampling

Sample for the Nutrition Dropoff consists of all refresher and panel respondents who completed the main interview in the wave 11 of SHARE survey in the Czech Republic.

Advance Letter/Invitation

The nutrition dropoff is not communicated in the advance/invitation procedures.

The consent for completing the nutrition dropoff will be obtained at the end of main interview. The questionnaires will be presented to the respondent immediately after the end of completion of the main interview.

Incentives: Payments for Respondents

In the information about the nutrition study, respondents are informed about an incentive payment for completing each nutrition dropoff Questionnaire: for the FFQ the incentive amount is 300 CZK, for the 24HDR both weekday and weekend questionnaires the incentive payment is 500 CZK. Therefore, the maximum amount of incentive payments for one respondent is 800 CZK. Because all respondents who complete the CAPI main SHARE interview will have already provided their payment information needed to receive the incentive payment, the additional tasks related to nutrition dropoff will consist of transferring the payment to a respondent’s bank account or by postal voucher together with the main interview completion incentives.

Payments for Interviewers

The Survey Agency is responsible for remuneration of survey agency coordinators and interviewers.

Fieldwork

After the end of main interview, the interviewer presents to the respondent the questionnaires and information about the Nutrition Dropoff.

FFQ

The FFQ questionnaire is an 8-page, self-completed, paper-and-pencil questionnaire left with the respondent. The FFQ questionnaire has a unique identification number ffqid written on the cover page of the questionnaire. Additional matching identifiers filled on each FFQ questionnaire will be the first name, year of birth, and gender of the respondent.

24HDR

The 24HDR questionnaires consists of a pair of 3-page memory cards to be filled with a delay for one weekday and one weekend day. The weekday questionnaire has a rpid identification number written on the top of the questionnaire, together with a day name and a date of working-day recall. The weekend questionnaire has a rvid identification number written on the top of the questionnaire, together with a day name and a date of weekend-day recall. Additional matching identifiers filled on each recall questionnaire will be the first name, year of birth, and gender of the respondent. The dates of weekday and weekend recall will be filled by the interviewer and approved by the respondent. The weekend recall will be scheduled at least after 7 days after the weekday recall questionnaire.

Each respondent will have two options how to complete the recall questionnaires: the respondent can use a web-tool application Intake24 developed by the MRS Unit at Cambridge University, or the respondent can send the filled memory cards to the survey agency in a pre-paid envelope. For the self-administered web option, each respondent will receive instructions how to connect to a website administered by the Survey Agency. At this website, the respondent will use the rpid and rvid identifiers to login to the Intake24 application that will guide the respondents to fill the recall questionnaires. The advantage of the self-administered Intake24 application is the photo gallery of food and drinks and is preferred from the point of view of data quality. After the completion of the recall questionnaire, the Intake24 software also offers the respondent to download immediately calculated summary dietary report as a pdf file.

The Intake24 software installed at the Survey Agency server also provides detailed information on the fieldwork progress. If the respondent does not complete a recall questionnaire within ten days of the agreed date of the recall, the survey agency will call the respondent and offer assistance with filling the questionnaire. The interviewer will inform the respondent that a telephone support will be available for filling the recall questionnaires.

6. Coding Guidelines

Coding of the FFQ and 24HDR dropoff questionnaires will be conducted on the secure server at the survey agency where the data will be stored and nutrition coding software installed. Both FETA and Intake24 are open-source free software, without any license or other fees. The survey agency will be responsible for the setup of the secure server, installation, and implementation of the software. All these tasks will be guided, approved, and monitored by SHARE-CZ. The MRC Unit team will provide support and assistance with installation of the software at the survey agency secure server. Costs related to the translation of the food database and cost of the MRC Unit support will be paid by SHARE-CZ.

Data coding of collected nutrition dropoff questionnaires will follow the guidelines and rules provided by SHARE-CZ based on the MRC Unit methodology. The MRC Unit provides clearly documented guidelines for all food categories (Data entry template, Milks look up list, Cereal look up list, Fats look up list). Survey agency will be responsible for coding and entering dropoff questionnaire data into the coding software. Methodology, guidance, teaching, and monitoring for this process will be provided by SHARE-CZ.

Coding of FFQ

For each respondent identified by pidcom, the survey agency will generate a ffqid identifier for inputting data from the FFQ questionnaire into the FFQ tool. Using the ffqid together with the matching identifiers (first name, gender, year of birth) the survey agency will code the information from the FFQ questionnaires into a csv file.

The FFQ data are coded automatically by an open source FETA software written in C and C++ languages, installed at the survey agency secure server. FEAT software inputs the comma-separated data file from the collected data. The FETA software is free, open source, fully downloadable from its official site. In FETA, frequency category is converted into a portion multiplier. After multiplication with the portion size, an average daily food weight for each of the FFQ items is obtained. These weights are multiplied by the nutrient composition per gram to obtain the nutrient composition of the actual amount eaten. After summing all FFQ items for a respondent, an average daily nutrient intake is obtained. It is recommended that individuals with more than 10 missing lines of data are excluded. The top and bottom 0.5% of the ratio of energy intake to estimated basal metabolic rate are flagged as extreme outliers of nutrient intakes.

Coding of 24HDR

For each respondent identified by pidcom, the survey agency will generate one rpid identifier for a weekday recall questionnaire and one identifier rvid identifier for a weekend-day recall questionnaire. Using the rpid and rvid together with the matching identifiers (first name, gender, year of birth), the data from memory cards will be inputted into the Intake24 tool.

If the respondent opted for using the web-based Intake24 application, the data are directly entered into the database. If the respondent sent the filled memory cards by mail to the survey agency, the recall information is entered by an operator at the Survey Agency into the web-based application on behalf of the respondent.

For the 24HDR questionnaires, the survey agency will follow all technical procedures for localization documented at Intake24 documentation space and wiki pages. Intake24 tools and datasets for translation and harmonization, and the source code are hosted on GitHub under the terms of the Apache2.0 license and can be viewed, forked, and adapted. Intake24 provides detailed research guides, including food database system for local adaptation, including food images, food and drink sizes, and all other tools necessary for localization.

7. Nutrition dropoff data management

MRC Unit support and access

  1. The MRC Unit at University of Cambridge will provide the open-source food database, software for coding the FFQ and Intake 24 recall questionnaire, and will provide support for installing the software at the survey agency secure server.

  2. The MRC Unit will not have access to the secure server at the survey agency, will not have access to the collected data nor personal information of respondents.

  3. The MRC Unit researchers will have the same access to the nutrition data as any other registered user only at the SHARE-ERIC public release database.

Data storage and coding

  1. The collected data will be stored in an RSA-encrypted form at the survey agency secure server. The survey agency is responsible for ensuring that the collected data are always separate from contact information and any other SHARE and personal data.

  2. The Survey Agency will be responsible for coding the collected dropoff questionnaires. The coded data will be used as input data files for the FETA and Intake24 software. Data coding will be monitored by the SHARE-CZ team.

  3. When coding and the FETA and Intake24 software analysis is complete, the survey agency will transfer a RSA-encrypted nutrition dropoff questionnaire data to SHARE-CZ secure server using the HTTPS transfer protocol. The data file will not contain any contact information or other SHARE data. The datafile will contain only data acquired from the FFQ and 24HDR questionnaires, their software generated nutrition values, pidcom and matching identifiers (first name, gender, year of birth).

  4. The SHARE-CZ team will create the final nutrition dropoff Data File by analyzing the data using scientifically validated nutrition methods developed. The analysis will not use any personal or contact information of the respondents.

Delivery of the nutrition dropoff Data file to SHARE-ERIC

  1. SHARE-CZ team will upload the nutrition dropoff Data File to the secure central SHARE server (SHARE Transfer server, which is run by SHARE Central and uses the HTTPS protocol) together with full documentation.

  2. The SHARE Central team will merge the delivered data with the SHARE main data using the common identifiers. This data will be made available to registered scientific researchers in accordance with the SHARE Conditions of Use. Only this released data may be used for research subsequently. The usage of any internal data for research purposes is prohibited.

8. Licenses, materials, safety and interviewer training

Licenses

Both FFQ and Intake24 questionnaires are open source questionnaires. The FETA and Intake24 software is freely available for research use with acknowledgment of the EPIC-Norfolk study and Cambridge MRC Epidemiology Unit. Additional costs related to MRC Unit team support for software installation and localization will be paid by SHARE-CZ. SHARE-CZ is responsible for coordination and collaboration with the FFQ and Intake24 providers.

Ethical review

There are no health or any risks associated with the nutrition dropoff. All questionnaires are validated, internationally accepted questionnaires on nutrition. There is no need for an ethical committee review.

Data ownership

Ownership of all data collected and generated by the nutrition dropoff will belong to SHARE-ERIC. The data will be linked to the main SHARE data and provided in the open-access regime to all registered users.

Materials

The survey agency will be responsible for the hardware and software needed for the project. All tasks will be guided and monitored by SHARE-CZ. SHARE-CZ is responsible for cooperation with the FFQ and Intake24 providers.

The survey agency will be responsible for the preparation and distribution of all printed materials (e.g. instructions, information and support materials, the questionnaires) for interviewers and respondents. Telephone support will be provided to assist respondents with completion of the questionnaires. The whole process will include functional testing, consistency checks and field rehearsal testing. The survey agency will ensure that the system complies with applicable data protection regulations.

Safety

There are no safety issues or risks associated with the nutrition dropoff.

Training of Interviewers

Personnel at the Survey Agency will be trained to follow the methodology and software installation needed for coding the collected nutrition data. SHARE-CZ will be responsible for coordinating the training.

Appendix 1: Nutrition Dropoff Questionnaire

see: sharew11-dropoff-cz-en.pdf